Survey of Income and Program Part~cipation

Residential Mobility of One-Person Households

No. 8803 50 James Witte and Herbert Lahmann

German Ins t i tu te for Economic Research

July 1988

Th i s paper was p resen ted a t t h e Census Bureau's Fou r t h Annual Research Conference h e l d March 20-23, 1988. a t t h e Na t i ona l C l a r i o n Ho te l i n A r l i n g t o n , V i r g i n i a . The views expressed i n t h i s paper a r e t h e a u t h o r ' s and do no t n e c e s s a r i l y r e f l e c t t h e views o f t h e Bureau of t h e Census.

PREFACE

The i nc rease i n t h e number o f persons l i v i n g a lone and t h e p a t t e r n s o f r e s i d e n t i a1 mobi 1 i t y among one-person households can on l y be unders tood i n 1 i ght o f t h e t r a n s i t i o n s between one-person and mu1 t i -person house- ho lds. The ' l i ke l ihood- o f a change i n d w e l l i n g o r household compos i t i on v a r i e s among one-person households, most obv ious l y w i t h t h e age of t h e Person 11 v i n g alone. T h i s paper p resen ts e x p l o r a t o r y research on one- person households, t h e processes l e a d i n g t o t h e i r f o rma t i on and d i s - s o l u t i o n and t h e r e s i d e n t i a1 rnobi l i t y assoc ia ted w i t h such households. The a n a l y s i s i s based on panel da ta c o l l e c t e d over a two y e a r t i m e p e r i o d i n t h e U n i t e d S ta tes and t h e Federa l Repub l i c o f Germany.

TABLE OF CONTENTS

PREFACE

1 INTRODUCTION ....... ....... ......................................... 1

2 SOURCES OF DATA AND ANALYTICAL FRAMEWORK....,...................... 2

2.1 The Ana l ys i s o f Comparative Panel Data ................... .... 2

... 2.2 Res iden t i a1 M o b i l i t y and Cahnges i n Household Composition.. 5

3 EMPIRICAL RESULTS .................................................. 7

3.1 Immobi 1 e One-Person Households, .............................. 7

3.2 Pure R e s i d e n t i a l M o b i l i t y ..................................... 9

3.3 The Foundat ion o f One-Person Households....................... 14

.................. 3.4 The T r a n s i t i o n t o a M i l t i - P e r s o n Household.. 16

3.5 The I n d i v i d u a l Components o f Aggregate Change i n t h e Number o f One-Person Households............................... 22

3.6 Model 1 i ng Resi d e n t i a1 Mobi 1 i t y o f One-Person Households i n t h e FRG ................ ........................ 23

4 CONCLUSIONS ........................................................ 29

REFERENCES ............................................................ 31

1 INTRODUCTION

The cu r ren t popu la t ion o f one-person households i s made up of persons o f d i f f e r e n t ages, a t vary ing stages i n t h e i r f am i l y careers. Young persons who have l e f t t h e parenta l household bu t have not y e t formed a f a m i l y of t h e i r own are one s i g p i f i c a n t component o f t h e popu la t ion of one-person households. Other persons 1 i v i ng alone a re somewhat o lder : persons who l e f t t h e parenta l household some t ime ago and never formed a fami ly of t h e i r own, as we1 1 as those who have on ly recen t l y begun l i v i n g a1 one. Some i n d i v i d u a l s i n t h e l a t t e r group cont inued t o l i v e w i t h t h e i r parents longer than most persons o f t h e i r generation. Others are divorced, widowed o r separated from t h e i r spouses, and s t i l l o thers began l i v i n g alone a f t e r having l i v e d w i t h o ther r e l a t i v e s o r non-re lated persons. The l a r g e s t b loc of one-person households a r e eadecly. &ny of these persons on l y began t o l i v e alone recent ly , a f t e r l ong years o f l i v i n g w i t h others, w h i l e a few have not l i v e d w i t h o ther persons s ince l eav ing t h e parenta l household l ong ago. The under ly ing premise of t h e research presented below i s t h a t t h e r e s i den t i a1 mobi 1 i t y o f one-person households i s inseparable from t h e processes by whi ch such households a re formed and d i ssol ved a t d i f f e r e n t stages i n t h e l i v e s o f i nd i v idua ls .

Moverover, along w i t h t h e d i f f e r e n t age r e l a t e d pa t te rns o f t r a n s i t i o n s between one-person households and mu1 t i -person households t h e economic forces of l a b o r and housing markets a l s o i n f l u e n c e t h e charac ter and frequency of r e s i d e n t i a1 mobi 1 i t y over t h e course o f i n d i v i d u a l s ' 1 ives. Young persons, re1 a t i ve ly f r e e from fami ly ob l i ga t i ons , a re i d e a l candidates f o r empl oyment-re1 ated r e s i d e n t i a l mob i l i t y . Once an i n d i v i d u a l has found a more o r l e s s s t a b l e p o s i t i o n i n the labor market, t h e dec is ion t o move i s more l i k e l y t o be a response t o t h e cos t and q u a l i t y o f housing. I n l a t e r years several con t rad i c to ry fac to rs come i n t o play. On t h e one hand, t h e i n d i v i d u a l , no longer bound by t h e r e s p o n s i b i l i t i e s o f employment, i s f ree t o move. I n add i t ion , f i n a n c i a l considerat ions o r phys ica l impai rment may fo rce t h e person t o f i n d smal ler quarters. On t h e o the r hand, t h e s o c i a l t i e s t o a p a r t i c u l a r p lace and dwel l ing, acccumulated du r ing t h e preceeding years o f r e l a t i v e i m n o b i l i t y , may ho ld over i n t o t h i s phase of a person's l i f e . I n short , observable pa t te rns of r e s i d e n t i a l m o b i l i t y a re t h e j o i n t product o f economic and soc ia l forces t h a t combine and i n t e r a c t i n d i f f e r e n t ways a t d i f f e r e n t phases o f people's l i v e s .

I n pa r t , our research concentrates on one-person households because t h e com- p l ex process o f household mobi lily f s . r i m p l i f ied when t h e household cons i s t s of one person r a t h e r t han many. More impor tan t l y , however, one-person house- ho lds have assumed a c r i t i c a l r o l e i n t h e o v e r a l l household s t r u c t u r e a t both t h e macro and micro leve ls . This can be seen i n t h e i nc reas ing number of persons 1 i v i n g alone. I n t h e pas t twen ty - f i ve years t h e number of one- person households i n t h e Federal Republic o f Germany (FRG) has more than doubled--from j u s t over 4 m i l l i o n i n 1961 t o 8.8 m i l l i o n i n 1985 ( S t a t i s t - i sches Bundesamt 1987 ). The p r o p o r t i o n o f one-person househol ds among a1 1 households i n t h e FRG thereby increased from 20.6% i n 1961 t o 33.6% i n 1985. I n t h e Un i ted States (US) t h e number o f one-person households rose even more d ramat i ca l l y from 6.9 m i l l i o n i n 1960 t o 20.6 m i l l i o n i n 1985 and t h e pro- p o r t i o n o f one-person households among a1 1 households i n t h e US i ncreased from 13.1% i n 1960 t o 23.7% i n 1985 (U.S. Bureau o f t h e Census 1987a). Th i s

. .

means t h a t i n a q u a r t e r o f a cen tu r y t h e p r o p o r t i o n o f one-person households i n t h e FRG i ncreased by 63.1% and i n t h e US by 80.9%; a vas t change i n l i v i n g arrangements over a p a r t i c u l a r l y s h o r t p e r i o d o f t ime. S i m i l a r r a t e s of change can be found- i n o t h e r European and Scandi nav i an c o u n t r i e s (Roussel 1983).

The i ncreas i ng p r o p o r t i on o f one-person househol ds among a1 1 households i s 1 i n k e d t o concre te eco-nomi c and techno1 og i c a l devel opments--i n p a r t i cu1 a r t h e r i s i n g s tandard o f l i v i n g , t h e improved income and l a b o r market p o s i t i o n o f women, and improved b i r t h c o n t r o l techniques. A t t h e same t i m e t h e r e has been a s h i f t i n popu la r values, norms and i d e a l s , which by no means a re u n r e l a ted t o changing economi c c i rcumstances. Independence, i nd i v i dua l freedom and s e l f - r e a l i z a t i o n a re t h e catchwords f o r t hese changes i n con- temporary c u l t u r e ' s normat i ve f ramework. Demographic a n a l y s i s o r econometr i c a n a l y s i s based on aggreyate data, o r models bu i l t around how i n d i v i d u a l s have ac ted i n t h e pas t a re n o t s u f f i c i e n t t o e x p l a i n t h i s change o r t o cons ide r i t s consequances. The i n c r e a s i n g p r o p o r t i o n o f one-person house- ho lds can on l y be ~ ~ n d e r s t o o d by c o n s i d e r i ng t h e changi ng behav io ra l p a t t e r n s o f i nd i v i duals.

Th i s papee desc r i bes recen t developments i n t h e popul a t i on o f one-person households i n t h e Federa l Repub l i c o f Germany and t h e U n i t e d States. I n t h e f o l l o w i n g s e c t i o n t h e da ta used i n t h e a n a l y s i s i s d iscussed and a framework f o r d e s c r i b i n g t h e co inc idence o f changes i n household composi t i on and r e s i d e n t i a1 mobi 1 i t y i s i n t roduced. Based on r e p r e s e n t a t i ve panel da ta on one-person households i n t h e US and t h e FRG, t h e e m p i r i c a l p a r t of t h e paper concent r a t e s on t hose households which acqui r e addi t i onal members o r change dwe l l i n g s and t h e e x t e n t t o which bo th t ypes o f change occur s imul taneous ly . The f o r m a t i o n o f new one-person households i s s i m i l a r l y d iscussed and t h e two processes a re t hen examined i n combinat ion t o cons ide r how t h e y i n t e r a c t t o determine t h e aggregate number o f one-person households i n each of t h e two coun t r i es . F i n a l l y , methods t o i n c l u d e t h e f o r m a t i on and d i s s o l u t i o n of one-person households i n models o f r e s i d e n t i a l m o b i l i t y a r e d iscussed and e x p l o r a t o r y r e s u l t s u s i n g a 1 og i t reg ress i on model a re presented.

2 SOURCES OF DATA AND ANALYTICAL FRAMEWORK

2.1 The Ana l ys i s o f Comparative Panel Data

Our research on one-person households i s based on two s e t s o f panel data: t h e German Sozi o-okonomisches Panel (SOP) and t h e 1984 Panel f r om t h e Census Bureau 's Survey o f Income and Program P a r t i c i p a t i o n (SIPP) (1) . These panel s t u d i e s a r e s i m i l a r i n many respects . Th i s f a c t g r e a t l y a i d s comparat ive research, f o r t h e r e i s always t h e r i s k t h a t d i f f e r e n c e s between two s o c i e t i e s a r e confounded w i t h d i f f e r e n c e s i n s tudy des ign and research methodology. Perhaps t h e most impo r tan t d i f f e r e n c e between t h e two s t u d i e s i s t h e s h o r t e r t i m e p e r i o d between SIPP waves, f o u r months as opposed t o t h e yea r - l ong i n t e r v a l between SOP waves. F o r some research ques t ions , f o r example, t h e exac t amount o f program b e n e f i t s r e c e i ved, more f r e q u e n t i n t e r v i ews may we11

--------- ( 1 ) An overv iew o f t h e SOP Panel may be found i n Hane fe ld (1984) and o f SIPP

i n U.S. Bureau o f t h e Census (1987b).

reduce r e c a l l e r ro r . However f o r something as s a l i e n t as changes i n house- h o l d composit ion, t h e longer t ime span between SOP waves i s of l ess s i g n i f i c - ance. Owing t o our concern f o r t h e mobi 1 i ty o f one-person households i t i s more important t h a t both panels use e s s e n t i a l l y t h e same f o l l o w i n g r u l e s (2).

The fo l l ow ing r u l e s adopted by both S I P P and SOP have one p a r t i c u l a r l y impor tan t imp1 i c a t i o n ' for t h e study o f one-person households. A dynamic approach i s necessary t o determine wh4ck households g i ve r i s e t o one-person households, as w e l l as those t h a t absorb persons who p rev ious l y l i v e d alone. Both panels share t h i s advantage and y e t b i n d i t by t h e same r e s t r i c t i o n s : persons not se lec ted i n t h e o r i g i n a l sample who subsequently en te r t h e sample (because they en te r a household w i t h an a d u l t member o f t h e o r i g i n a l sample) a re no longer in te rv iewed when they s top I f w i n g w i t h a member o f t h e o r i g i n a l samp'l e. This means, by d e f i ni t ' ron, t h a t a l l persons i n one-person households i n both s tud ies are members o f t h e o r i g i n a l sample. By l i m i t i n g one's ana lys is t o members o f t h e o r i g i n a l sample some of t h e more complicated problems rooted i n panel data, i n p a r t i c u l a r those r e l a t e d t o miss ing da ta and weight ing, are s i m p l i f i e d .

Our ana lys is o f German one-person households i s based on t h e f i r s t t h r e e waves of t h e SOP, i n te rv iews t h a t took p lace i n t h e s p r i n g o f 1984, 1985, and 1986. A one-person household i s de f ined as one i n which a person i s found t o be 1 i ving alone on a t l e a s t one o f t h e i n t e r v i e w dates. I n t h i s way a d i s c r e t e t ime s t r u c t u r e i s imposed on a process t h a t i s e s s e n t i a l l y continuous; j u s t because t h e i n te rv iews f a l l i n a regu la r cyc le does no t mean t h a t changes i n household composit ion f o l l o w t h e same rhythm. As a r e s u l t some in fo rmat ion concerning t h e exact t i m i n g o f changes i n household composit ion i s l o s t and s h o r t s p e l l s o f 1 i v i n g alone may go unobserved. As t h e purpose o f t h i s paper i s p r imar i l y exp lora tory , t h e dec i s ion was made t o s a c r i f i c e a measure of p r e c i s i o n t o s i m p l i f y t h e problem. However, i t should be emphasized t h a t both panels da te changes i n household membership, as we1 1 as a number of o ther processes, more prec ise ly . L a t e r work can adopt a more accurate and r e a l i s t i c t i m e s t ruc ture .

Given a d i s c r e t e t ime framework, t h e subsequent changes i n household s t r u c t u r e experienced by persons i n a panel a re considerably s i m p l i f i e d . At each o f t h e t h r e e po in t s , an i n d i v i d u a l belongs t o e i t h e r a one-person o r a mul t i -person household. Panel members i n one-person households i n 1984 ( T I ) f a l l i n t o one of f o u r ca tegor ies i n 1985 (T2) : 1 ) those con t i nu ing t o l i v e alone, 2 ) those now l i v i n g w i t h o ther persons i n a mul t i -person household, 3) those who have died, emigrated o r moved i n t o an i n s t i t u t i o n and thus l e f t t h e populat ion, 4 ) and those who have l e f t t h e panel (wave nonresponse), whereby t h e i r s t a t u s regard ing popu la t i on membershi i s open. The same f o u r possi b i 1 i t i e s e x i s t between 1985 (T2) and 1986 (737 . Other panel members i n mul t i -person house- holds i n 1984 ( T l ) o r 1985 (T2) stopped l i v i n g i n mul t i -person households between 1984 and 1985 o r between 1985 and 1986 and represent newly formed

--------- ( 2 ) I n a panel t h e f o l l o w i n g r u l e s a re j u s t as impor tan t as t h e o r i g i n a l

sample desi gn i t s e l f (Kal t o n and Lepkowski 1985). Based on t h e o r e t i c a l concerns (such as t h e boundaries o f t h e popu la t i on o f in ference) and p r a c t i c a l concerns (such as f o l l o w i n g mobi le members of t h e o r i g i n a l sample) t hey p rov ide an ongoing d e f i n i t i o n o f who w i l l be i n te rv iewed and t h e popu la t i on represented by t h e 1 ong i tud i na l sample.

one-person households. I n t h i s manner data i s a v a i l a b l e f o r persons con- t i n u a l l y l i v i n g i n one-person households du r ing t h i s t i m e per iod, as w e l l as f o r persons who have begun t o l i v e alone and those who have begun l i v i n g i n households w i t h o ther persons. Data concerning American one-person house- holds f rom t h e 1984 S I P P Panel was then organized i n a s i m i l a r manner (3).

Weight ing presents a problem f o r a l o n g i t u d i n a l ana lys i s of t h i s t ype us ing - t h e SIPP publ ic-use cross-sect ional f i l e s . The Census Bureau's recommenda- t i on i s t h a t t h e i n i t i a l Wave 1 weights may be used i f , as i s t h e case here, t h e ana lys i s i s s o l e l y based on members o f i n te rv iewed Wave 1 house- holds. However, t h e Census Bureau a l so st resses t h a t t h e weights do no t account f o r sample a t t r i t i o n (Census Bureau 1987b). I n t h e course o f devel opi ng 1 onyi t u d i na l weights f o r t h e SOP p r o j e c t i t has become apparent t h a t sample a t t r i t i o n i s associated w i t h changes i n household composit ion and r e s i d e n t i a l m o b i l i t y and t h a t t h e i r i n f l u e n c e i s p a r t i c u l a r l y s t rong i n t h e case o f one-person households (Rendtel 1987). To compensate f o r t h i s problem a s imple s t ra tegy proposed by L i t t l e and David (1983) was adopted. A l l persons i n one-person households a t T1 but no longer i n t h e panel a t T2 were i d e n t i f i e d and t h e reasons f o r t h e i r departure f rom t h e panel were explored us ing t h e publ ic-use c ross-sec t iona l f i l e s f o r t h e i n t e r v e n i n g t i m e period. Persons who l e f t t h e sample due t o death, emigra t ion o r i n s t i t u t i o n a l i z a t i o n were thus d i s t i ngu i shed f rom cases of nonresponse.

The T1 weights o f one-person households who remained i n t h e panel a t T2 were then weighted-up us ing an adjustment f a c t o r equal t o t h e i nve rse o f t h e response r a t e f o r one-persons households. L ikewise persons i n m u l t i - person households a t T1 who subsequently formed one person households, bu t then become nonrespondents by T2 were used t o ad jus t t h e weights o f persons who began t o l i v e alone a f t e r T1 and remained i n t h e panel a t T2. The same procedure was then performed t o f u r t h e r ad jus t t h e weights t o compensate f o r nonresponse between T2 and T3. The adjustment f a c t o r s f o r t h e T2 and T3 t ime pe r iod a l so inc luded those cases l eav ing t h e sample due t o t h e 15% reduct ion i n t h e SIPP sample t h a t took p lace between T2 and T3. Whi le a number o f more soph is t i ca ted techniques have been developed ( L i t t l e and David 1983; Ka l ton 1987; Rendtel 1988) these adjustment f ac to rs a re adequate f o r t h e exp lo ra to ry ana lys i s presented here and produce est imates of t h e popu la t i on o f one-person households very c lose t o c ross-sec t iona l est imates

--------- ( 3 ) Three t ime p o i nts--September 1983 (T I ) , September 1984 (T2) and September

1985 (T3)--were selected. The p u b l i c re lease cross-sect ional f i l e s f o r t h e f i r s t e i g h t panel waves were then read t o i d e n t i f y persons who l i v e d i n one-person households a t any t i m e between T1 and T3. Based on t h e respondent's r o t a t i on group t h e appropr i a t e waves and re ference months were then used t o determine t h e t ype of household t o which i n d i v i d u a l s belonged a t each o f these t i m e po in ts . Cases where t h e e n t i r e i n t e r v i e w was imputed f o r one o r more o f these t i m e p o i n t s were dropped, owing t o t h e p a r t i c u l a r problems associated w i t h imputed data i n analyses o f change. (Census Bureau, 1987b).

prepared by t h e Census Bureau us ing t h e S I P P data (4).

2.2 Resident i a1 Mobi li t y and Changes i n Household Composition

Panel s tud ies are designed t o measure change, but c o l l e c t i n g i n d i v idua l mobi 1 i t y data i s never' easy, even i n t h e context o f p r o j e c t s 1 i ke SIPP and SOP, where spec ia l e f f o r t s are made t o f o l l o w mobi le sample members. Some o f t h e problems a re concrete quest ions o f measurement and sample s e l e c t i o n and t h e r e l a t e d costs and benef i ts . I n add i t i on , complex a n a l y t i c a l and conceptual issues a re unavoidable i n t h e study o f changes i n household composition, sueh as t h e quest ion o f t h e proper operat ional i za t ion and d e f i n i t i o n o f h ~ u s e h o l d s and t h e r e l a t i onship o f i n d i v i d u a l s t o households over t i m e (McMi 1 l e n and Werr io t 1985; Duncan 1985). Resident i a1 mobi 1 i t y i s o f t e n a quest ion o f i n d i v i d u a l and no t household m o b i l i t y ; y e t i t i s a f fec ted by t h e s t a b i l i t y o r I n s t a b i l i t y a f household composition. Changes i n house- h o l d composit ion a l t e r housing needs and thus o f t e n l ead t o r e s i d e n t i a l mobi 1 i t y (5 ) .

F i yure 1 describes f i v e d i f f e r e n t degrees o f household m o b i l i t y ( A through E ) t h a t may r e s u l t from d i f f erent combi nat i ons o f r e s i d e n t i a1 mobi 1 i t y and changes i n household composition. Case A , immobi l i t y , i s t h e s implest , household composit ion remains unchanged and a l l persons remain t h e t h e same dwel l ing. case 0 , ure r e s i d e n t i a l m o b i l i t y , occurs when t h e household moves t o a d i f f e r e n t dwel!ing and t h e r e i s no change i n household composition. M o b i l i t y o f t h i s t ype i s t y p i c a l l y rooted e i t h e r i n t h e households l i v i n g s i t u a t i o n ( i n t h e dwe l l i ng o r t h e neighborhood) o r l abo r market f ac to rs , whereby one o r more members may be "pushed out" o f an area due t o poor l a b o r market cond i t ions , and/or " p u l l e d t o " another area due t o t h e prospects of a b e t t e r 1 abor market pos i t ion .

( 4 ) For example, t h e monthly average o f one-person households J u l y t o September 1983 i s given as 19.71 mi l l i o n (US Bureau o f t h e Census 1984) and f o r t h e same t ime pe r iod 1984 as 20.78 m i l l i o n (US Bureau of t h e Census 1985) an increase of 5.4% i n t h e number o f one-person households. The same r a t e o f increase between 1984 and 1985 would y i e l d 21.91 m i l l i o n one-person households and increase o f 11.2% over t h e two year period. Our r e s u l t s f o r September 1983 y i e l d 20.3 m i l l i o n one-person households and f o r September 1985 22.6 m i l l i o n - an increase o f 11.3%.

(5) These events need no t be simultaneous: a t i m e l a g i n e i t h e r d i r e c t i o n i s possib le. The e x i s t i n g members o f a household may move i n a n t i c i p a t i o n o f a change i n household composit ion o r a f t e r t h e change has occurred. Moverover, t h e percept ion of housing needs, as w e l l as t h e p o s s i b i l i t y t o f u l f u l perceived needs, a re sub jec t t o t h e f u l l gamut o f s o c i a l and economic cons t ra i n t s and necessi t ies.

F i gure 1: Re1 a t i uonshi ps between Changes i n Household Composition and Resi dent i a1 Mobi 1 i t y

Resi dent i a1 E f f e c t on Case Change i n Household Mobi 1 i t y Type of one-person

Composi t i on Yes No M o b i l i t y household

A No change X Immobi l i t y No change

B No change

C 1 I n d i v i duals en ter house- ho ld ( i n c l u d i n g b i r t h s )

C2 I n d i v i d u a l s leavehouse- ho ld ( i n c l u d i n g death and erni g r a t i on)

Pure New dwel l i ng r e s i dent i a1

X Latent T r a n s i t i o n t o mu1 t i -purpose household

X Latent O i ssol u t i on

C 3 I n d i v i d u a l s en ter and leave X La ten t D i s s o l u t i o n

D l I n d i v i d u a l s en ter house- X Complex T r a n s i t i o n t o ho ld ( i nc l udi ng b i r t h s ) mu1 t i -person

household

D 2 I n d i v i duals leave house- X ho ld ( i n c l u d i n g death and emigrat ion)

Compl ex D i ssol u t i on

D3 I n d i v i d u a l s en ter and leave X Comp 1 ex Di ssol u t i on

E A1 1 i nd i v i duals 1 eave - - Household D i s s o l u t i o n popul a t i on d i s s o l u t i o n

Cases C 1 through C3 r e s u l t when t h e r e i s a change i n household composit ion-- t h e a d d i t i o n of new members and/or t h e l oss o f o l d members--whereby one o r more o l d members o f t h e household remains i n t h e dwel l ing. We r e f e r t o such changes as l a t e n t m o b i l i t y , s ince a d d i t i o n o r l o s s o f members o f t e n r e s u l t s i n changed housing needs but a move may not be desi r a b l e o r poss ib le a t t h e time. A move may subsequently f o l l o w o r may have p res ious l y occurred, if t h e change i n household composi t i on was ant ic ipa ted . Cases D l through D3 i nvol ve t h e same types o f changes i n household composit ion, bu t a re coupled w i t h r e s i d e n t i a l m o b i l i t y and are re fe r red t o as complex m o b i l i t y . E i t h e r t h e move o r t h e change may p r e c i p i t a t e t h e other : l a r g e r o r smal le r quar te rs may be found t o ad jus t t o t h e change i n household composition. The move t o smal ler quar te rs may f o r c e someone t o leave t h e housheold; o r a l a r g e r dwe l l i ng may a1 1 ow o r necess i ta te t h e a d d i t i o n of household members. Empi r i ca l l y t h e d i s t i n c t i on between 1 a ten t and compl ex mobi 1 i t y depends on how c lose together a change i n household composit ion and dwe l l i ng need t o be i n order t o be de f ined as simultaneous. For t h e ana lys is presented below a y e a r l y cyc le has been chosen: complex mobi 1 i t y means t h a t bo th househol d composit ion and d w e l l i n g change between T 1 and T2 o r between T2 and T3. An impor tan t quest ion t o consider i s t h e ex ten t t o which cases o f pure r e s i d e n t i a l m o b i l i t y are 1 agged responses t o o r a n t i c i p a t i ons of changes i n household composit ion.

The f i n a l t ype of mob i - l i t y , household d i s s o l u t i o n (Case E ) , occurs when a l l members of t h e household simultaneously leave t h e dwe l l i ng and are no longer p a r t of t h e populat ion. Theoret ica l l y household d i s s o l u t i o n on ly occurs upon t h e death of a1 1 household members, bu t i n empi r ica l research t h e inc idence of household d i s s o l u t i o n i s determined by the f o l l o w i n g r u l e s and d e f i n i t i o n of t h e popu la t ion o f in ference. I n t h e S I P P and SOP s tud ies household d i s s o l - u t i o n occurs through t h e i n s t i t u t i o n a l i z a t i o n , emigra t ion o r death of a l l househol d members.

One-person households are an appropr ia te s t a r t i n g p o i n t f o r t h e study of t h e re1 a t i onship between changes i n household composit ion and r e s i den t i a1 mobi 1 i t y , because t h e d i f f e r e n t types o f m o b i l i t y are l o g i c a l l y s i m p l i f i e d . The p o s s i b i l i t i e s C2, C3, D2 and D3 e x i s t only f o r mul t i -person households; f o r w i t h one-person households, household d i s s o l u t i o n r e s u l t s when t h e one and on l y member leaves t h e household. The th ree types o f l a t e n t m o b i l i t y (Cl, C2, and C 3 ) and complex m o b i l i t y (Dl , D2, and D3) each reduce t o a s i n g l e form (when a person enters t h e household), t he t r a n s i t i o n t o a mult i -person household. The household does not leave t h e populat ion, bu t ceases t o e x i s t as a one-person household. As a case o f l a t e n t m o b i l i t y , t h e t r a n s i t i o n t o a mu1 t i - pe rson household means t h a t t h e i n d i v i d u a l i nvo l ved has remained a t h i s o r her o l d address and a d d i t i o n a l persons have j o ined t h e household. I n t h e event o f complex m o b i l i t y e i t h e r t h e i n d i v i d u a l has moved i n w i t h another person o r persons, who remain a t t h e i r o l d address, o r both p a r t i e s change addresses and together occupy a common dwe l l i ng a t a new address.

I n add i t i on , t h e study o f one-person households w i t h i n a panel design touches on t h e important bu t d i f f i c u l t quest ion o f household formation. Household formation i s i n e x t r i c a b l y t i e d t o t h e processes described as l a t e n t and com- p l e x m o b i l i t y . The departure o f a household member from any t ype o f household and t h e subsequent format ion o f a one-person household by t h i s person i s a case o f t h e fo rmat ion o f a one-person household through complex m o b i l i t y . A c h i l d l eav ing t h e parenta l dwe l l i ng t o e s t a b l i s h h i s o r her own household i s perhaps t h e most simple example o f t h i s t ype o f complex m o b i l i t y . The fo rmat ion o f a one-person household takes t h e form o f l a t e n t m o b i l i t y when a1 1 bu t one member of t h e household leave t h e dwel l ing. Those l e a v i n g t h e dwel l i n g may form one o r more one-person households o f t h e i r own, w h i l e t h e i ndi v i dual remaining i n t h e o l d dwe l l i ng d e f a u l t s t o a one-person household, as l ong as no a d d i t i o n a l persons occupy t h e dwel l ing.

3 EMPIRICAL RESULTS

3.1 Immobile One-Person Households

P rev i ous research i n d i c a t e s a number of associat ions between l i v i n g alone and pa t te rns o r c l u s t e r s o f socio-economi c t r a i t s . The sex-speci f i c d i f f e r e n c e i n l i f e expectancy and t h e lower average marr iage age o f women leaves a number of women alone a t t h e end o f marr ied l i f e i n many i n d u s t r i a l coun t r i es (Spiegel 1987). Th is p a t t e r n i s f u r t h e r strengthened i n t h e FRG due t o t h e great number o f women who were d ivorced o r widowed du r ing t h e war years, a t ime when t h e remarr iage ra tes f o r women were much lower than today ( W i t t e 1988). Table 1 presents u n i v a r i a t e d i s t r i b u t i ons f o r several soc i o-economic va r iab les (age, sex, mar l t a l s ta tus , employment s ta tus , income and m i n o r i t y s t a t u s ) f o r immobi l e one-person households i n t h e FRG and t h e US between 1983 and 1986. The immobile households, persons 1 i v i n g a lone f o r t h e e n t i r e observat ion

Socio-economic Panel (SOP) Survey of Income and Program Participation (SIPP) - 1984 Panel

54 years 34 to 53 16 to 33 54 years 34 to 53 16 to 3 3 and older years years and older years years

female male female male female male female male female male female male

I ' in percent in percent

Age (-ZOO*) 60.7 10.1 8.2 7.7 7.1 6.2 53.4 14.8 8.7 10.4 5.6 7.1 Minority status 2)

white (-100%) - 9 - - - - 55.1 14.6 8.4 9.7 5.5 6.7 non-w. (-100%) - - - - - 41.5 16.2 11.1 15.1 6 . 3 9.8

Marital mtatus 3) married, living apart 0.8 single 10.9 divorced 6.9 widwed 67.0

Employment 8tatus 3) amp1 oyed 11.1 not employed 76.4

Income 3) lower tarcile 31.7 riddle tercile 34.0 ppper tercile 19.9

. Weighted population (thousands) 4,106 918 Observed canes 4 17 119

1) That is, without changetn in hounehold composition or dwelling. 2) Alien residents in the PRG are excluded I ' f- this table. 3) Based on the start of the observation period.

I Tha data in based on waves 1 - 3 of the SOP and on waves 1 - 8 of the 1984 SIPP panel public release files. Table 1

Boefo-eaonomio Cb8ractaristicm of One-Person Housebolds 1~0bil0 1) Households TI - T3

p e r i o d w i thout an address change, correspond t o t h e t y p i c a l image o f e l d e r l y one-person households. Women 54 years o f age and o l d e r make up over 60% of t h e irrmobi l e one-person households i n t h e FRG and 53% i n t h e US. If one considers male one-person households i n t h i s age group as we1 1, approximately 70% of t h e t o t a l number o f immobi l e one-person households i n both coun t r i es are made up o f persons 54 years o f age o r older. C h a r a c t e r i s t i c a l l y these persons a r e no t membefs o f t h e l abo r force--14% o f t h e German respondents were employed a t t h e beginning o f t h e observat ion pe r iod as compared t o 22% o f these American one-person households. Desp i te t h e 1 ower employment r a t e i n t h e FRG, o l d e r immobile one-person households i n t h e FRG are r e l a t i v e l y b e t t e r - o f f t han t h e i r American counterparts. Using t e r c i l e s based on monthly income o f a l l one-person households i n each country, j u s t about a t h i r d of a l l German but over 40% of American households i n t h i s age group f a l l i n t o t h e lowest income group.

I n t h e two younger age groups i n bo th count r ies more than 80% o f t h e persons i n immobi 1 e one-person households are employed. The 1 owest empl oyment quota i s found among t h e youngest German one-person households o f t h i s type, which i s presumably expla ined by t h e longer p e r i o d o f educat ion i n t h e FRG. S i m i - 1 a r l y whi 1 e over ha1 f o f t h e youngest American one-person households f a1 1 i n t o t h e h ighes t income category on ly 40% o f Germans i n t h i s group have t h i s l e v e l o f income. On t h e o ther hand, among persons 34 t o 53 years of age, 64% o f Germans i n imrnobi l e one-person households have incomes i n t h e upper t e r c i l e as compared t o 59% o f t h e Americans.

Perhaps t h e most s t r i k i n g d i f fe rence, however, between t h e imnob i le one- person households i n t h e two younger age groups i s t o be found i n t h e i r m a r i t a l s tatus. Over 90% o f t h e persons i n t h e youngest immobile one- person households i n t h e FRG have never been marr ied, as compared t o 76% i n t h e US. Correspondingly on ly 6% of t h e Germans i n t h i s group had been divorced, as compared t o 18% of t h e Americans. Among those o l d e r t han 34 and younger than 54 years o f age, d i vo rce i s more common among persons i n immobi l e one-person households i n t h e FRG (27%). Nonetheless a f a r . g rea ter p r o p o r t i on o f persons i n t h i s group i n t h e US have been d ivorced (39%).

3.2 Pure Res iden t i a l M o b i l i t y

Cases o f pure r e s i d e n t i a l m o b i l i t y a re a subset o f those persons who l i v e d a1 one f o r t h e e n t i r e observat ion period. Pure r e s i den t i a1 mobi 1 i t y occurs when persons move i n t h e absence of a change o f household composit ion. There i s l i t t l e d i f f e r e n c e between t h e US and t h e FRG i n t h e p ropo r t i on of cases o f r e s i d e n t i a l m o b i l i t y t o t h e t o t a l popu la t i on o f s t a b l e one- person households--12.5% i n t h e FRG and 14.6% i n t h e US. I n both countr ies, as Table 2 shows, m o b i l i t y o f t h i s t y p e i s c l e a r l y associated w i t h age and gender. As a r u l e , moving f rom t h e o ldes t t o t h e youngest age group t h e p r o p o r t i on o f persons changi ng addresses i ncreases , whereby w i t h i n each age group men a r e more mobi le than women. The gender-speci f ic d i f f e r e n c e i s weaker i n t h e US, no t on ly i n t h e middle age group (where women a re s l i g h t l y more mobi le t han men), bu t a l so i n t h e o ldes t and t h e youngest age categories. I n t h e FRG t h e gender-speci f ic d i f f e r e n c e decl ines, moving from t h e o ldes t t o t h e youngest age groups, bu t even among one-person households younger than 34 men a r e s t i l l move l i k e l y t o change addresses than women. Otherwise i n t h e FRG, w i t h i n t h e i n d i v i d u a l age groups, t h e r e i s no th ing i n t h e socio- economic c h a r a c t e r i s t i c s o f those who move t h a t se ts them apar t f rom t h e

54 years 34 to 53 16 to 33 pnd older Yelrs YBCLtE) Total

rn Total 7.2

Male 12.9 Female 6.2

Total 8.2 14.4 38.0 14.6

Male 12.3 Female 6.9

Based on waves 1-3 of the SOP and waves 1-8 of the SIPP 1984 panel public release files. (FGR: n = 661, estimated weighted population = 6.1 million; US: n = 2,604, estimated weighted population - 14.4 million)

T8bl. 2 Mobility of Stable One-Person Households According to Age Percentage Moved in tho Coursq of 2 Yoar Obaorvatfon Poriod

immobile one-person households. S i m i l a r l y i n t h e US, apart f rom age and gender, t h e cases o f pure r e s i d e n t i a l m o b i l i t y do not d r a s t i c a l l y d i f f e r from t h e immobile one-person households. I n t h e youngest age group, s t a b l e one-person households who moved are l ess l i k e l y t o be d ivorced and are l ess we1 l - o f f f i nanci a1 l y than those who remained i n t h e same dwe l l i ng du r ing t h e observat ion period. These d i f fe rences, however, can be t raced back t o an age d i f f e r e n c e between t h e two groups (6).

Extensive in format ion regarding housing qua1 i t y , s a t i s f a c t i o n w i t h housing and housing costs i s an important element o f t h e SOP data. The hypothesis t h a t pure r e s i d e n t i a l m o b i l i t y o f one-person households i s connected p r i n c i p a l l y w i t h t h e i r housing s i t u a t i o n may be more c l o s e l y examined i n t h e l i g h t o f t h i s data. To begin wi th, those who move a re quest ioned regarding t h e reason f o r t h e move. Table 3 presents t h e reasons g iven by a l l household reference persons, regardless o f household size, and compares these w i t h t h e reasons g iven i n t h e cases o f pure r e s i d e n t i a l m o b i l i t y of one-person households. For t h e l a t t e r group t h e responses a re f u r t h e r broken down according t o age. Among a l l household reference persons t h e move t o a new dwe l l i ng i s most o f t e n (42%) r e l a t e d t o housing concerns. A s l i g h t l y g rea te r p ropor t i on (45%) o f t h e persons 1 i v i ng a1 one a t t r i b u t e d t h e move t o such concerns--i n p a r t i c u l a r t h e des i re f o r a dwe l l i ng o f t h e appropr ia te s ize, more comfort, a b e t t e r l o c a t i o n o r ren t , or t h e purchase

(6) The mean age among t h e movers i s 36, two years younger than those remaining a t t h e same address (T-value = 7.35 w i t h 605 d f ). Th is d i f f e r e n c e f i t s w i t h t h e r e s u l t s presented below and i n sec t i on 3.6 where we f i n d t h a t i n t h e youngest age group i n t h e FRG, m o b i l i t y i s most common among those not y e t s e t t l e d i n t o occupational careers.

o f a house o r condimini urn. Among a l l households changing addresses, 12% of t h e household reference persons described t h e move as occupation o r employ- ment re la ted , as compard t o 19% o f t h e one-person households. Near ly one t h i r d o f a l l moves were described as fami ly - re la ted . However, as one would expect, such reasons are f a r l ess common f o r s t a b l e one-person households (13%). Presumably these persons moved t o be c lose r t o f a m i l y members i n o ther households o r t h e move was a lagged response t o a change i n household composit ion t h a t took p lace p r i o r t o t h e observat ion period. When one con- s ide rs t h e t h r e e age groups separate ly , t h e importance o f occupat i ona1 p lace- ment f o r r e s i d e n t i a l m o b i l i t y among young persons becomes apparent. Housing re1 ated reasons were named l e a s t commonly (39%) and occupational and employ- ment r e l a t e d reasons most commonly (26%) among t h e youngest age group (6).

Improvement Asked to Employment Family of dwelling leave by related related ~ituation landlord X U $ - TeCLsOn8 Q&hSX

in percent

One-Person Households 45.4 5.5 18.8 13.1 17.2

Age > 54 years 41.5

55-34 years 67.1 33-16 years 39.2

All Households 41.8 5.9 12.2 32.3 7.9

Based on waves 1-3 of the SOP (observed cases = 826 including 92 one-person households)

Tabla 3 Roaaona Oivon for Moving in tho FRO

All Rousaholda Cornparad to 8tablo Ono-pormon Bousaholds

A d i f f i c u l t y w i t h t h e design o f t h i s quest ion i s t h a t i t fo rces respondents t o name a s i n g l e most important reason f o r t h e move and thereby obscures t h e f a c t t h a t a v a r i e t y of f a c t o r s may be involved. Th is becomes apparent when one compares t h e c h a r a c t e r i s t i c s o f prev ious dwel l ings w i t h those o f new dwel l ings (see Table 4). The housing s i t u a t i o n o f one-person households i n t h e youngest age group i s genera1 l y improved through a move. Thus t h e s i z e of t h e d w e l l i n g and t h e number o f rooms increased. The average r e n t a l s o increased considerably, whereby i n most cases t h e increased r e n t i s deemed appropriate. S u r p r i s i n g l y t h e assessment o f t h e need f o r renovat ion of t h e b u i l d i n g i n which t h e dwe l l i ng i s l oca ted i n d i c a t e s a d e c l i n e i n

(6 ) The mean age among t h e movers i s 36, two years younger than those remaining a t t h e same address (T-value = 7.35 w i t h 605 d f ) . Th is d i f f e r e n c e f i t s w i t h t h e r e s u l t s presented below and i n sec t i on 3.6 where we f i n d t h a t i n t h e youngest age group i n t h e FRG, m o b i l i t y i s most common among those not y e t s e t t l e d i n t o occupational careers.

In the event of pure residential mobility 1) 1984 1986 34-53 16-33 2 54 34-53 years years years years

in percent

In the event of immobility 2)

16-3 3 years

> 54 34-53 16-33 - years years years

in percent

> 54 - years

Need for renovation of the building completely somewhat not at all

Facilities with bath, with central heating with bath, without central heating without bath, with central heating without bath, without central heating

Number of rooms 1

2 3 4 and more

Size of dwelling m2 0 - 40 41 - 60 61 - 80 81 and more

Rent. (DM) 1 - 200

201 - 300 301 - 400 401 and more

Satisfaction with housing costs favourable appropriate too high

Satisfaction with housing very matisfid matisf ied dissatisfied

The data im based on waves 1 - 3 of SOP. 1) Those who moved; observed cases n = 92 (estimated weighted popula- tion = 826 thousands). 2 ) Those who remained at the same address; observed cases n = 661 (estimated weighted population = 5,795 thousands).

Table 4 Changm in Dwelling Characteristics of One-Person Households

housing qual i t y , i .e., t he p ropo r t i on o f respondents desc r ib ing t h e b u i l d i n g as i n need o f renovat i on increased a f t e r t h e move. This d e c l i n e i n more general i n d i c a t o r s of housing qual i t y ( c o n d i t i o n o f t h e b u i l d i n g , i n c l u d i n g hea t i ng and bath f a c i 1 i t i e s ) i n d i c a t e s t h a t an improvement i n housing q u a l i t y i s no t t h e p r i n c i p a l aim o f r e s i d e n t i a l m o b i l i t y among persons i n t h i s age group.

On t h e o ther hand, persons i n t h e o ldes t age group who moved and l i v e d alone f o r t h e e n t i r e observat ion p e r i o d thereby a t t a i n e d an improvement i n t h e i r housing according t o these general i n d i c a t o r s o f housing q u a l i t y , 89% o f these persons judged t h e b u i l d i n g s i n which t h e i r new dwel l ings were loca ted as no t i n need o f renovation--as compared t o on ly 69% o f persons i n t h e youngest age group. L ikewise persons i n t h e o ldes t age group improved t h e f a c i l i t i e s o f t h e i r dwe l l ings as a r e s u l t o f moving. 95% o f these persons had c e n t r a l heat ing and a bath o r shower i n t h e i r own dwel l ings, as compared t o 74% o f those i n t h e youngest age group. Based on these measures of housing qual i t y , persons i n t h e youngest age group acqui red dwel l ings w i t h f a c i l i t i e s comparable t o t h e average f o r t h e FRG, w h i l e those i n t h e o ldes t age group acqui red b e t t e r than average housing (7).

Over 75% o f a1 1 stab1 e o lde r one-person households occupied dwel l ings w i t h two o r t h r e e rooms i n a d d i t i o n t o t h e k i t c h e n and 20% l i v e d i n one-room apartments. Only h a l f t h e youngest one-person households occupied dwel l ings w i t h two o r t h r e e rooms, w h i l e a t h i r d occupied one-room apartments. Among those who moved, those i n t h e youngest age group tended t o move t o l a r g e r dwel l ings and those i n t h e o ldes t t o smal ler dwel l ings. Economic considera- t i o n s most 1 i k e l y p layed a r o l e here; as those moving t o smal le r dwe l l ings were p r i m a r i l y widows w i t h low monthly incomes.

On t h e whole, persons i n t h e o ldes t age group p a i d h igher ren ts a f t e r moving than beforehand. Apparently f o r these persons a bet ter-equi pped dwel l i ng was wor th t h e p r i ce ; t h e m a j o r i t y o f those who moved judged t h e i r new r e n t t o be appropriate. Su rp r i s i ng l y , however, t h e p ropo r t i on o f persons i n t h i s age group who judged t h e i r r e n t as 'much t o o h igh" was not decreased through r e s i d e n t i a l m o b i l i t y . Presumably t h i s i s r e l a t e d t o t h e supply of a f fo rab le smal l e r dwel l i ngs. Studies have shown t h a t d i s s a t i s f a c t i o n w i t h t h e r e n t f o r a dwe l l i ng i s p r i m a r i l y a judgement of t h e r e n t i n r e l a t i o n t o a v a i l a b l e income (Lahmann 1988). Persons who f i n d t h e i r r e n t t o be "much t o o h igh" a re concentrated i n t h e 1 owest income category and among those persons not i n t h e l a b o r force. Also among t h e youngest age group t h e p r o p o r t i o n of persons who judge t h e i r r e n t t o be excessive i s r e l a t i v e l y h i g h (9%) owing t o t h e low incomes a v a i l a b l e t o some o f these households.

--------- ( 7 ) In t h i s contex t i t should be mentioned t h a t a wel l-equipped d w e l l i n g i s

a necessary p r e r e q u i s i t e f o r t h e ex is tence o f a one-person household. These persons need a techn ica l l y we1 1 -equipped and e f f i c i en t dwel l i ng t o reduce t h e e f f o r t associated w i t h housework and thereby acqu i re t h e t i m e necessary f o r l a b o r market and rec rea t i ona l a c t i v i t i e s .

14

Stable one-person households older than 33 and younger than 54 years of ageare general ly comprised of persons who have found their p lace on the labormarket. Str ick data conf ident ia l i ty procedures preclude the use of the SOPdata for research regarding the distances and places associated with resid-ent ia l mobi l i ty . However recent research using other data has shown thatthe rate of employment-related resident ia l mobi l i ty is greatest between theages of 20 and 24- and thereafter steadily decreases (t{agner 1987). Keepingin mind that the analysis here is l imi ted to households wi thout changes inhousehold composi t ion, one would expect that one-person households in th isage group pr imari ly move to improve their housing si tuat ion. t lore than20% of the persons in th is age group said the reason for thejr move wasthe purchase of a house or codominium. l , loverover,939 of the respondentsin th is group were sat isf ied wi th their housing af ter the moYe. The highleve l o f sa t is fac t ion is a lso l inked to an inprovement in the hous ingqual i ty (according to object ive measures such as heat ing and bath faci l i -t ies ) as we l l as an inc rease in the average dwel l ing s ize . The averagerent has increased as wel l , however i t is general ly v iewed as an appropr iateincrease. Among the stable one-person households in th is age group there isa c lear di f ference in housing qual i ty between those who nove and those whoremain in the same dwel l ing. Even though relat ively high incomes are foundamong both groups those who move tend to occupy lower qual i ty housing evenafter the move than those who remain in the same dwel l ing.

3.3 The Foundat ion of One-Person Households

The f i rst d ist inct ion to consider by the format ion of one-p€t lson householdsis whether or not the process is one of complex or latent nobi l i ty , that ls ,whether or not the person forming the household changes drel l ings. Complexmobility is the more cormon type of formation of one-person households inboth the FRG and the US. In the FRG 67tr of the one-person households formedduring the observat ion per iod involved complex mobi l i ty , as compared to 62Sin the US (see Table 5). As wi th pure resident ia l nrobi l i ty among one-personhouseholds, the formation of one-person households through complex nobil itymost f requent ly involves persons in the youngest age group: ln the FRGnearly 901 of the persons who formed one-person households {n this mannerwere younger than 34, in the US 75f of these persons rere ln th is age grouP.The formation of one-person households through latent rather than complexmobi l i ty is more cormon in the oldest age group-- in the FRG and in the USnearly half (48%) of those persons who formed one-person hwseholds in thlsmanner were older than 54 years of age.

The clearest difference between the FRG and the US is the greater proportionof one-person households formed by divorce in the US. The fornation of one-person households through complex mobil ity among the youngeat age group andthe formation of one-person households through latent mobi"l lty in the oldestage group contain the largest number of cases and are most suited for a compar-ison between the two countries. In both instances the proportion of divorcedpersons ls greater in the US than ln the FRG.

IIIItI

Complex Mobility Latent Mobility Socio-economic Panel (SOP)

> 54 34-53 16-33 2 54 34-53 16-33 - years years years years years years

in percent Age (-100%) 2.3 8.3 89.4 48.8 32.0 19.2 Minority status

white (-1001) non-w. (-1001)

Sex male 12.1 70.0 55.0 33.6 77.5 54.2 female 87 9 30.0 45.0 66.4 22.5 45.8

Marital status 2) married, living apart 12.1 single - divorced 23.7 widowed 64 2

Employment status 2) employed 34 . 4 not employed 65.6

Income 2) lower tercile 64.2 middle tercile 35.8 upper tercile

Complex Mobility Latent Mobility nram Survey of Income and Pro,

Participation (SIPP) - 1984 Panel > 54 34-53 16-33 > 54 34-53 16-33 - -

years years years years years years in percent

7.2 17.4 75.4 48.7 22.7 28.6

I Weighted population (thousands) 35 i 129 1,378 370 243 14 5 365 888 3,840 1,528 7 12 897 Observed cases 6' 17 133 47 25 2 6 76 192 799 352 165 194

1) Formation is the transition from a multi-person household to a one-person household. This transition is considered complex mobility when it occurs in conjunction with a change in address and latent mobility when the person remains at the same address. 2) Marital status after the change in household composition is used. Income and employment status are based on the start of the observation. Alien residents in the FRG are excluded from this table.

I The data im based on waveo 1 - 3 of the SOP and on waves 1 - 8 of the SIPP panel public release files. Tabla 5

locio-eaonomic Characteristics of One-Person Rouseholds Household Formation 1)

The s o c i a l processes commonly l ead ing t o t h e fo rmat ion o f one-person house- holds--chi l d r e n l eav ing t h e parenta l household, d i vo rce and death of a spouse--vary i n t h e i r r e l a t i v e importance between t h e FRG and t h e US. Th i s may be seen by cons ider ing t h e t ype o f household i n d i v i d u a l s l i v e d i n p r i o r t o t h e formation o f a one-person household and changes i n m a r i t a l s ta tus t h a t accompany t h e t r a n s i t i o n f rom a mul t i -person t o a one-person household (see Table 63. Over 40% o f t h e one-person households formed i n t h e FRG between 1984 and 1986 were es tab l ished by c h i l d r e n l e a v i n g t h e parenta l household. I n t h e US t h i s process accounts f o r on ly a qua r te r of a l l new one-person households. Death o f a spouse a l s o p lays a more important r o l e i n t h e fo rmat ion o f one-person households i n t h e FRG than i n t h e US--widows and widowers made up 15% o f t h e new one-person households i n t h e FRG as compared t o 10% i n t h e US.

I n t h e US, on t h e o ther hand, d ivorce, separat ion and t h e d i s s o l u t i o n of households compri sed of two un re la ted persons were more f requen t l y associated w i t h t h e format ion of one-person households than i n t h e FKG. The departure of a d i s t a n t r e l a t i v e o f t h e head o f household (a person o ther than t h e c h i l d o r spouse o f t h e head of household) accounts f o r on ly 5% o f t h e newly formed one-person households i n t h e FRG. I n t h e US, however, 21% o f t h e one-person households were formed by a more d i s t a n t r e l a t i v e o f t h e head of household. Some o f these cases may i n v o l v e t h e departure o f c h i l d r e n brought i n t o t h e household by t h e cu r ren t pa r tne r o f t h e head o f household. This would account f o r some o f t h e d i f f e rences between t h e US and t h e FRG i n t h e p ropo r t i on o f one-person households created by c h i 1 dren 1 eavi ng t h e parenta l household. However, on ly 32% o f these " d i s t a n t re1 a t i ves " were under t h e age o f 34 as compared t o 92% o f c h i l d r e n who l e f t t h e parenta l household. F i n a l l y , a 1 arger p ropo r t i on o f t h e one-person household i n t h e FRG (17%) were formed by persons l eav ing households t h a t contained more than one o the r unre la ted person than i n t h e US (7%). The p ropo r t i on o f one-person house- holds formed by each o f these processes and t h e manner i n which t h i s va r i es between t h e FRG and t h e US c l a r i f i e s which processes are more o r l e s s import - an t f o r t h e format ion o f one-person households i n each country. One should a1 so consider t h e s i gni f i cance o f t h e newly formed one-person households i n r e l a t i o n t o t h e number o f one-person households e x i s t i n g a t t h e s t a r t of t h e observat ion period. The f i n a l two columns o f Table 6 presents t h e est imated number o f households formed i n t h i s manner as a percentage of t h e t o t a l est imated number o f oneperson households a t t h e s t a r t of t h e observat ion period. Taki ng a1 1 o f these processes together , t o t a l number o f one-person households formed i n t h e FRG dur ing t h e obser3atf on pe r iod equals 30% o f t h e number a t t h e s t a r t of t h e observat ion per iod, w h i l e i n t h e US t h i s p r o p o r t i o n amounts t o over 40%. While t h e number of one-person households i n each country grew considerably du r ing t h i s t i m e per iod, t h e r a t e o f growth was by no means t h i s h igh and was, i n f a c t , g rea ter i n t h e FRG than i n t h e US. However, t h e i nc reas ing number o f one-p-erson households i s no t s imply a f u n c t i o n o f t h e fo rmat ion o f one-person households. To understand t h i s process one must consider t h e d i s s o l u t i o n of one-person households and t r a n s i t i o n s t o m u l t i -person households as we1 1.

3.4 The T r a n s i t i o n t o a Mult i -Person Household

A t t h i s p o i n t i t i s important t o r e c a l l t h e d i s t i n c t i o n made above between t h e t r a n s i t i on o f one-person households i n t o mu1 ti -person households and household d i sso lu t i on , when a1 1 members o f t h e households 1 eave t h e popul at ion.

The socio-economic c h a r a c t e r i s t i c s of persons i n t h e FRG and t h e US who stopped l i v i n g alone and formed a household w i t h o the r persons a re presented i n Table 7. I n t h e case o f complex m o b i l i t y t r a n s i t i o n s t o mul t i -person households predomi nan t l y i nvol ve persons i n t h e youngest age categ0r.y i n both countr ies. I n t h e event o f l a t e n t m o b i l i t y , whereby a one-person household stays a t t h e same address bu t becomes a mul t i -person household, i n both count r ies a g reater p ropo r t i on o f persons i n t h e two o l d e r age groups a r e involved. I n t h e US over 30% o f t h e i n d i v i d u a l s who stopped l i v i n g alone and remained a t t h e same address were over t h e age o f 53.

As w i t h t h e formation o f one-person households, t h e t r a n s i t i o n t o m u l t i - person households may occur through one o f several d i s t i n g u i s h a b l e s o c i a l processes. A person l i v i n g alone may marry, a pe r iod o f m a r i t a l separat ion may end, a c h i l d may r e t u r n t o t h e parenta l household, o r a person may begin l i v i n g w i t h o the r unre la ted persons. I n Table 8 f o u r processes l ead ing t o t h e t r a n s i t i o n t o a mult i -person household a re described and f o r each country t h e p r o p o r t i on o f changed one-person households a t t r i b u t a b l e t o each i s presented. The i n d i v i d u a l s m a r i t a l s ta tus be fore and a f t e r t h e t r a n s i t i o n . as w e l l as t h e i n d i v i d u a l ' s r e l a t i o n s h i p t o t h e head o f household and o the r members of t h e household were used t o ca tegor ize t h e observed t r a n s i t i o n of one-person households.

I n both count r ies approximately one- th i rd o f a l l instances o f t r a n s i t i o n t o a m u l t i -person household took p lace through marr iage o r t h e end o f a pe r iod of m a r i t a l separation. I n t h e US, 36% o f t h e t r a n s i t i o n s t o mul t i -person households i nvo l ved t h e i n d i v i d u a l forming a household w i t h re1 a t i v e s o the r than h i s o r her spouse, whereby over one-thi r d o f these i nvo l ved c h i l d r e n r e t u r n i n g t o t h e parenta l household (8). I n t h e FRG t h e p ropo r t i on of one-person households d isso lved through t h e format ion o f a mul t i -person household w i t h r e l a t i v e s o the r than one's spouse i s on ly h a l f as large. The p r i n c i p l e process l ead ing t o t h e t r a n s i t i o n t o a mul t i -person household i n t h e FRG was t h e format ion o f a two person household w i t h a person (not necessar i l y o f t h e o the r sex) no t r e l a t e d by blood o r marriage--45% of a l l t r a n s i t i o n s t o mul t i -person households. I n more than 75% o f these cases t h e persons described t h e i r r e l a t i o n s h i p as an unmarried couple ("Lebensgeme- inschaf t " ) . Changes o f t h i s t ype were l e s s o f t e n associated w i t h t h e t rans - i t i o n t o a mul t i -person household i n t h e US (25%). I n bo th count r ies t h e number of persons l i v i n g alone who subsequently formed a household w i t h more than one o the r un re la ted i n d i v i d u a l on ly p lays a minor ro le.

--------- (8) It may be assumed t h a t an even greater p r o p o r t i o n o f these cases i nvo l ved

t h e r e t u r n o f a c h i l d t o t h e parenta l household, t h a t i s , when t h e one- person household being observed was t h e parent and t h e new household member was t h e ch i l d . Unfor tunate ly t h e da ta se ts used f o r t h i s ana lys i s were n o t s t r u c t u r e d so as t o e a s i l y i d e n t i f y changes i n t h i s type.

Process leading to One-person households Weighted number of households the formation of a formed during the observation formed relative to weighted one-person household period total of one-person households

at the start of the observation period 4)

FRG US FRG US

in percent

Child leaving parental household 1)

Separation or divorce 12.2 17.9 3.6 7.2

Death of spouse 14.9 9.8 4.4 3.9

Person leaving household containing two unrelated persons 10.0 19.4

Person leaving household with related family members 2) 5.1 21.5

Parson leaving household with more than two unrelated persons 3) 16.8 6.8

Total Obsented cases

1) Only when one-person household formed by the child. Cases where a child mwes out and a remaining single pa- rent thus becomes a one-person household are treated as "Person leaving household with related family membersn. 2) The person leaving is related to one or more persons who remain, but is not the child or spouse of the head of household. 3) The person leaving is related to no one in the household though they may be related to one another. 4) The total number of weighted households formed is slightly, less here than in Table 9 due to 5 % of the cases in the FRG and 15 in the US where the process leading to formation could not be precisely identified.

The data i a based on waves 1 - 3 of the SOP and waves 1 - 8 of the 1984 SIPP panel public release files. Table 6

Proceases mading to the Formation of One-Person Rouseholds in the Federal ~epublic of Germany (PRO) and the United Btates (US)

Socio-economic Panel (SOP)

Complex nobility Latent Mobility 5 4 34-53 16-33 - - > 54 34-53 16-33

years years years years years years in percent

Age (-100%) - 18.9 81 .1 14.7 22.3 63.0 Minority status

white (-100%) - - - - - - non-w. (-100%) - - - - - -

Sex - male 75.7 50.3 31.2 ' 42.2 42.4 female - 24.3 49.7 68.8 57.8 57.6

Marital status 2) married, living together - 49.6 27.6 4.8 20.2 43.2 married, living apart o .. 1.4 - 4.2 - single 9.6 69.0 10.9 19 .1 52.7 divorced - 20.7 2.0 41.3 4 . 1 widowed 20.1 - 84.2 15.1 -

Eaployment atatus 2) employed - 100.0 96.3 5.9 100.0 . 73.8 not employed o - 3.7 49.1 - 26.2

Income 2) Aower tercile - 28.2 10.9 16.5 33.9 middle tercile - 30.0 21.3 53.2 4.2 26.2 upper tercile - 70.0 50.5 35.9 79.3 39.9

Weighted population (thousands) o 64 277 62 94 2 67 Observed cases o 12 4 1 9 15 3 9

Survey of Income and Program . Participation (SIPP) - 1984 Panel

Complex Mobility Latent Mobility > 54 34-53 16-33 > 54 34-53 16-33 - -

years years years years years years in percent

11.5 16.3 72.2 30.2 29.8 40.0

1) This transition is considered complex nobility when it occurs in conjunction with a change in address and latent mobility when the person remains at the same address. 2) Marital status after the change in household composition is used. Income and employment status are based on the start of the observation. Alien residents in the FRG are excluded from this table.

The data is based on waves 1 - 3 of the SOP and on waves 1 - 8 of the SIPP panel public release files.

Table 7 socio-economic Characteristics of One-Person Households

Transition to a Multi-Person Household 1)

Table 8 a l s o describes t h e importance of each o f these processes o f t r a n s i t i o n t o mu1 t i -person households re1 a t i ve t o t h e number o f one-person households a t t h e s t a r t o f t h e observat ion period. Viewed from t h i s perspect ive. each of these types^ o f t r a n s i t i o n i s r e l a t i v e l y more important i n t h e US than i n t h e FRG. Taken as a whole. t h e weighted est imate o f t h e number o f one-person household i s d isso lved do t o t h e t r a n s i t i o n from a one-person t o a m u l t i - person household equal t o 24% o f t h e weighted est imate of one-person house- holds i n t h e US a t t h e s t a r t o f t h e observat ion period. This p ropo r t i on i s considerably smal le r i n t h e FRG. amounting t o j u s t over 10% o f t h e t o t a l number o f one-person households i n 1984.

F i n a l l y . one must consider persons l i v i n g alone who l e f t t h e popu la t ion du r ing t h e observat ion period. Death and i n s t i t u t i o n a l i z a t i o n are t h e most common causes o f household d i s s o l u t i o n and, as one would expect. household d i s s o l u t i o n occurs most f requen t l y i n t h e o ldes t age group i n both countr ies. For t h e purposes a t hand. household d i s s o l u t i o n i s most important as a f u r t h e r component o f t h e aggregate number o f one-person households. I n t h i s regard household d i s s o l u t i o n i s of r e l a t i v e l y equal importance i n t h e two count r ies : t h e weighted est imate of t h e number o f one-person households l e a v i n g t h e popu la t ion i n t h e FRG amounts t o 6%. and i n t h e US 5%, of t h e weighted est imate o f t h e t o t a l number o f one-person households a t t h e s t a r t of t h e observat ion period.

Process leading to the transition to a multi-person household

One-person households Weighted number of households changed during the changed 2) relative to weighted observation period total of one-person households

at the start of the observation period

FRG US FRG US

in percent

Uarriage or end of marital separation

Formation of household with other family members 3) 17.7 35.8

Formation of household with one unrelated person 45.5 25.4

Formation of household with more than one unrelated person 4.4 6.3

Total Observed cases

1) ~bcludes only cases of dissolution of one-person households, when the individual forms a household with other persons and not the dissolution of the entire household, i.e., where a person living alone leaves the po- pulation through emigration, institutionalization or death. 2) The total number of weighted households dissol- ved is far lower here than in Table 9, which includes cases of household dissolution. 3) None of whom is the spouse of the person previously living alone.

The data is based on waves 1 - 3 of the SOP and waves 1 - 8 of the 1984 SIPP panel public release files. Table 8

Processes Loading to the Transition to a Multi-Person Household in the Federal Republia of Germany (PRO) and the United Btates (US)

3.5 The I n d i v i d u a l Components o f Aggregate Change i n t h e Number o f One-Person Households

Using t h e r e s u l t s presented above i t i s poss ib le t o descr ibe t h e manner i n which t h e growth i n t h e number o f one-person households has taken p lace i n t h e Un i ted States and t h e Federal Republ ic o f Germany (see Table 9). The i nc reas ing number o f ' one-person households can be decomposed i n t o var ious components: an e x i s t i n g stock o f one-person households, minus those one- person households d isso lved through death, emigrat ion, i n s t i t u t i o n a l i z a t i o n o r t r a n s i t i o n t o a mu1 t i - pe rson household, p l u s newly formed one-person households.

One-Person Households Tl 7 891 20,344

New One-Person Households Between T1 and T3 + 2,464 Dissolution of One-Person Households Between T1 and ~ 3 * m One Person Households T3

Percentage newly formed T1 - T3 relative to T1 Percentage dissolved T1 - T3 relative to Ti Growth in number of One-person Households T3 relative to T1

* includes persons leaving the population and transitions to multi-person households.

'P8b1.9 Change8 in the Population of One-Paraon Hou8ahold8

in tha Fadaral Rapublio of oarmany and tha Unitad Statam

I n t h e FRG t h e o r i g i n a l stock o f 7.9 m i l l i o n one-person households i n 1984 grew t o 8.9 m i l l i o n i n 1986. Th is i s t h e end r e s u l t o f processes o f house- ho l d f ormati on, d i ssol u t i on and t rans format ion i nvol v i ng a f a r g rea te r number o f households. Nearly 2.5 m i l l i o n new one-person households were formed du r ing t h i s t ime, w h i l e an a d d i t i o n a l 1.4 m i l l i o n were d issolved: t h e persons l i v i n g i n these households e i t h e r began li v i n g w i t h o ther persons o r 1 e f t t h e populat ion.

The 14% growth i n t h e number of one-person households du r ing t h i s t i m e p e r i o d i n t h e Federal Republ ic o f Germany d i d no t s imply r e s u l t f rom t h e format ion o f t h i r t e e n new one-person households f o r each hundred a l ready ex i s t i ng . Ins tead f o r each hundred e x i s t i n g one-person households t h i r ty new one-person households were formed, w h i l e seventeen, some o f which were j u s t formed w i t h i n t h i s t i m e per iod, were dissolved.

I n the US the growth i n the number o f one-person households between 1983 and 1985 was somewhat slower (11%), but the c i r c u l a t i o n o f i nd iv idua ls w i t h i n the popul at i on of one-person households was re1 a t i vely h i gher. The number o f -new one-person households formed dur ing t h i s per iod (8.2 m i 11 i on ) amounts t o over 40% o f t he t o t a l number o f one-person households a t the s t a r t o f t he t ime period. However the l a rge r number o f newly formed one- person households i n the US d i d not produce a greater growth i n the number o f one-person households than i n the FRG, as the number o f one-person house- holds dissolved during t h i s t ime per iod was also r e l a t i v e l y higher i n the US. The p ropor t i on o f one-person households 1 eavi ng the populat ion was roughly the same i n both countr ies r e l a t i v e t o t he number o f one-person households a t the s t a r t o f t h e observation period, 5% i n the US and 6% i n the FRG. However, the number of one-person households t h a t became m u l t i - person households re1 a t i ve t o the number o f one-person households a t t he outset i s much greater, 24% i n t he US and 10% i n the FRG.

3.6 Model 1 i ng Residenti a1 Mobi 1 i t y o f One-Person Households i n the FRG

A wide var ie ty o f techniques have been used t o develop models o f r es i den t i a l mob i l i t y (see Rima and van W i ssen 1987, f o r an overview o f models developed i n a number of countr ies). A c ruc i a l element o f a l l such models are estimates of the t r a n s i t i o n rates between states--such as marriedlnot married, same dwell i nglnew dwell i ng or 1 i v i ng w i t h parents/not l i v i n g w i t h parents. Even the most sophist icated model i s 1 i m i t ed by t he accuracy w i t h which these t r a n s i t i o n rates are estimated. The previous descr ip t i ve analysis of one- person households i n the US and t he FRG imply two general considerat ions t h a t are necessary f o r an adequate estimate o f these t r a n s i t i o n rates f o r one-person households. F i r s t , the model must address changes i n household composition, f o r these are o f ten associated w i th a change i n dwellings. Secondly, one-person households cons t i t u t e a heterogeneous popul a t i on; a mode1 o f res iden t ia l mob i l i t y must e x p l i c i t l y accomodate t h i s var iety. One-person households vary, most obviously w i t h the age o f t he person l i v i n g alone, as t o the frequency o f changes i n household composition and dwelling. One-person households a lso d i f f e r as t o t he circumstances under which a person began l i v i n g alone and t he l f k e l i hood t h a t t h e person w i l l l a t e r begin l i v i n g w i t h other persons. F ina l l y , they vary i n t he f i t between t he current dwel l ing and the ind iv idua l along w i t h h i s o r her prospects f o r cont inuing t o l i v e alone.

The s t a t i s t i c a l techniques o f ten re fe r red t o i n t he soc ia l sciences as event- h i s to ry analysis seem especi a1 ly appropriate f o r t h e analysis o f res iden t i a1 m o b i l i t y because these models e x p l i c i t l y address the durat ion spent i n a p a r t i c u l a r state, such as t he occupancy o f a p a r t i c u l a r dwell ing, and t he in f luence of var iables associated w i t h leaving a state. An app l i ca t ion o f t h i s c lass of models t o t he quest i on o f res iden t ia l mob i l i t y i n t h e FRG can be found i n Wagner (1987). This study i l l u s t r a t e s not only t h e strength o f these methods but a lso the problems associated w i t h t h i s approach fo r t h e res iden t i a1 mobi 1 i ty o f one-person households. Wagner's study i s based on re t rospect ive 1 l f e h i s t o r y data, i nc lud ing migrat ion and mar i t a l h i s to r ies . As such i t concentrates on changes i n fami l y s ta tus and dwel l ings as reported by a representat ive sample o f set ected b i r t h cohorts--persons a1 1 born before 1952. However when one considers t h a t t h e most mobi l e one-person households belong t o t he group of persons aged 33 t o 16 i n 1984, i.e., born

i n 1951 o r l a t e r , i t becomes apparent t h a t a p a r t i c u l a r l y s a l i e n t p a r t of t h e cu r ren t popu la t i on o f one-person households i s excluded f rom t h i s data.

The SOP and SIPP data i n c l u d e event -h is to ry data, such as m a r i t a l and f e r t i l - i t y h i s t o r i e s , f o r a l l respondents; bu t t h i s data, too, i s inadequate. The c r u c i a l problem i s t h a t t h e fo rmat ion o f one-person households and t h e t rans - i t i o n t o mul t i -person households o f t e n f a i l t o co inc ide w i t h t h e cata logue of events inc luded i n these event -h is to r ies . Marr iage accounts, f o r example, f o r on ly about one - th i rd o f t h e t r a n s i t i o n s t o mul t i -person households i n t h e FRG and t h e US. Periods of l i v i n g w i t h an unre la ted person p l a y a c r u c i a1 r o l e i n t h e format ion and d i s s o l u t i o n o f one-person households i n both count r ies , y e t such changes i n household composit ion c o n s t i t u t e non- events i n t h e r e t rospect i ve components o f these data sets. Ret rospect i ve data of t h i s s o r t i s a r i c h supplement t o a t r a d i t i o n a l panel design, none- the less i t does not p rov ide t h e i n fo rma t ion necessary f o r t h e study o f s p e l l s o f 1 i ving alone p r i o r t o t h e s t a r t o f t h e observat ion period. Even i f one conf ines t h e ana lys is t o one-person households e x i s t i n g a t t h e s t a r t of t h e panel, t h e problem o f l e f t - c e n s o r i n g ( t h e i n a b i l i t y t o accura te ly da te t h e fo rmat ion o f one-person households a l ready e x i s t i n g a t T I ) poses a s e r i ous p rob l em f o r t h e appl i cat i on o f event-hi s t o r y techniques (9).

Analyses o f t h i s s o r t f o r t h e FRG w i l l f i r s t be poss ib le when a s u f f i c i e n t number o f s p e l l s o f 1 i ving alone have begun du r ing t h e course o f t h e SOP. Owing t o i t s l a r g e r sample s ize, analyses o f t h i s s o r t may be poss ib le w i t h t h e SIPP data, though t h e panel 's design a l lows on ly f o r t h e ana lys is o f very sho r t per iods o f l i v i n g alone. On t h e o ther hand, f o r t h e ana lys i s o f one-person households i n t h e US, t h e Panel Study o f Income Dynamics (PSID) represents a valuable supplement t o SIPP, as t h i s panel can prov ide da ta on very long s p e l l s o f 1 i v ing alone. An example o f t h e use o f p ropo r t i ona l and nonproport i onal hazard r a t e models t o evaluate t h e importance of mortgage ra tes on household m o b i l i t y us ing da ta from t h e PSID i s found i n Qu ig ley (1987).

For these reasons another approach was needed t o exp lore our hypothesis t h a t t h e determinants o f r e s i den t i a1 nobi 1 i ty o f one-person households vary according t o t h e t ype o f one-person household under considerat ion. For t h i s ana lys i s we r e s t r i c t e d our sample t o persons l i v i n g alone a t t h e s t a r t of t h e panel and def ined t h e outcome o f i n t e r e s t as a dichotomy: moved i n t h e nex t two yearslremained a t t h e same address. Rima and van Wissen (1987) use a l o g i t regression model as one component o f t h e i r "dynamic household r e l o c a t i o n model" f o r t h e c i t y o f Amsterdam. They d e r i v e c o e f f i c i e n t s f o r measures o f change i n household composit ion and character1 s t i c s of t h e dwel l i ng as p r e d i c t o r s o f a households " w i 11 i ngness-to-move. ' They use a d i f f e r e n t outcome v a r i a b l e ( w i 11 ingness t o move/unwi lli ng t o move) than ours (moved/remained a t t h e same address), because on l y c ross-sec t iona l da ta was ava i lab le . However, those households t h a t had moved i n t h e year p r i o r t o t h e i n t e r v i e w were coded as w i l l i n g t o move ( regardless o f t h e i r s t a t e d

--------- (9) Even under t h e most he r ioc of assumptions t h e r e t r o s p e c t i v e da ta a v a i l a b l e

i n t h e SOP data prov ides i n s u f f i c i e n t i n f o r m a t i o n t o adequately da te t h e fo rmat ion o f a s i zeab le o r t i o n (13%) o f t h e one-person households e x i s t i n g P a t t h e s t a r t o f t h e pane .

preference a t t h e t ime o f t h e i n t e r v i e w ) , *...because then t h e e f f e c t of household, changes i n t h e past year on t h e w i l l i n g n e s s t o move i s measured more prec ise ly . " As independent var iab les they used: age of t h e head of household, household s i z e a t t ime t, change o f household s i z e between t ime t and t ime t + l dwe l l i ng s i z e (number o f rooms), tenure s ta tus o f t h e dwe l l i ng (owner l ren ter ) and dwe l l i ng t y p e i n t h e SOP data prov ides (apar tment ls ing le fami ly u n i t ) . They deScribe t h e i r r e s u l t s as f i t t i n g t h e data w e l l and cap tu r i ng most o f t h e var iance i n t h e dependent var iab le . A1 1 o f t h e i r coef f i c i ents were " s i gni f i cant l y d i f f erent f rom zeroH and have t h e expected s ign, except f o r t h e i r i n d i c a t o r o f changes i n household s t r u c t u r e (an increase i n household s i z e ) . Due t o t h e s i m i l a r i t y i n t h e dependent v a r i - ables and t h e i n c l u s i o n o f household s i z e among t h e independent va r i ab les and t h e i n c l us i on o f household s i ze among t h e i ndependent v a r i abl es, t h i s model appears as a reasonable base model f o r consi d e r i ng r e s i den t i a1 mobi 1 i t y among one-person households us ing t h e SOP data (10).

Resu l ts f o r a mode1 i n c l u d i n g t h e main e f f e c t s f o r t h r e e independent v a r i - ables are presented i n column 1 of Table 10. The r e l a t i v e l y l a r g e s i z e and negat ive s ign o f t h e c o e f f i c i e n t f o r t h e constant r e f l e c t s t h e general l ack of r e s i den t i a1 mobi 1 i t y found among one-person households. The posi t i ve c o e f f i c i e n t s f o r a change i n household composit ion and a smal l number o f rooms i n a dwe l l i ng i n d i c a t e t h a t these f a c t o r s a re associated w i t h a change o f address, whi 1 e t h e negat i ve c o e f f i c i e n t s attached t o persons i n 1 arge dwel l ings and i n t h e two o ldes t age categor ies i n d i c a t e a greater l i k e l i h o o d t h a t these persons w i l l remain a t t h e same address. The o v e r a l l f i t of t h e model i s very good. I n f a c t t h e f i t i s so good t h a t a t t h e .05 l e v e l of s i g n i f i c a n c e one can not r e j e c t t h e hypothesis t h a t t h e expected l o g i t s produced by t h e model a re d i f f e r e n t than those associated w i t h a f u l l y sa tura ted model, a model i n c l u d i n g a1 1 possi b l e i n t e r a c t i o n s between t h e i ndependent v a r i abl es.

The normal s t ra tegy pursued i n t h e case o f a l o g i t model t h a t f i t s so w e l l i s t o consider i f any o f t h e parameters i n t h e model can be e l im ina ted w i thou t a s i g n i f i c a n t decrease i n t h e o v e r a l l f i t o f t h e model. The d i f fe rence between two models can be t e s t e d by s u b t r a c t i n g t h e L~ values f o r t h e two models. The r e s u l t i n g sum i s d i s t r i b u t e d approximately as a chi-square value w i t h df equal t o t h e d i f f e r e n c e i n d f ' s between t h e two models (Knoke and Burke, 1980). Columns 2 through 4 i n Table 10 present t h e r e s u l t s obta ined when each of t h e independent va r i ab les i s e l im ina ted from t h e model. I n no case i s t h e r e s u l t i n g increase i n t h e n mber o f degrees o f freedom s u f f i c i e n t compensation f o r t h e increase I n t h e LB r a t i o t o j u s t i f y exc lud ing a parameter from t h e o r i g i n a l model.

--------- (10) We have e l im ina ted two o f t h e va r iab les used by Rima and van Wissen f rom

our model because i n t h e case of one-person households they a r e so c l o s e l y c o r r e l a t e d w i t h o ther i ndependent va r i ab les : owners a re almost exc l us1 ve l y found i n t h e o ldes t age group, w h i l e a l l s i n g l e f a m i l y u n i t s f a l l i n t o t h e 1 argest category o f dwel l i ngs occupied by one-person households. Addi ng these va r iab les does no th ing t o enhance t h e model and s imply m u l t i p l i e s t h e number o f c e l l s conta i n i ng few observa t i ons.

Constant

Change from one-person household to multi- person household -307 - .245 .493

( eO72) 0 ( ,068) (. 066)

Number of rooms in dwelling

1-2

Age older than 53

Likelihood ratio - chi-square 5.319 23.297 20.929 54.574

df 12 13 14 14

Difference relative to model 1

Chi-square df

Based on Waves 1 - 3 of the SOP. N = 997: indicates a difference from model 1 significant at the ,001 level. (standard orrors)

Tablo 10 Logit Model8 with Dependent Varirblor .Yovod/Remainad at B u o Addresr

One-Paraon Houraholdr of All Ages in tha TRQ

Based on our d e s c r i p t i v e ana lys is o f one-person households and t h e v a r i e t y o f types o f persons l i v i n g alone, i t appears somewhat imp laus ib le t h a t such a model, desp i te i t s good s t a t i s t i c a l p roper t i es , i s an adequate represent- a t i o n o f t h e process o f r e s i d e n t i a l m o b i l i t y f o r one-person households. It seems q u i t e l i k e l y t h a t t h i s mode1 f i t s w e l l f o r t h e e n t i r e popu la t ion o f one-person households but i s inadequate f o r p a r t i c u l a r subgroups o f t h i s popul a t i on. To consi der t h i s possi b i 1 i t y t h e age group ca tegor i es used above present a rough but simple ca tegor i za t i on o f one-person households. Considerable heterogenei ty remai ns w i t h i n each age group. A more p r e c i se ana lys i s would c a l l f o r t h e o r e t i c a l l y more i n t e r e s t i n g categor ies and a more soph is t i ca ted and e f f i c i e n t o rgan iza t ion o f t h e data regarding t h e t i m i n g o f changes. The i n t e n t here i s no t t o produce a s ing le , "bestH model f o r each o f these age groups. Rather, we hope t o demonstrate t h a t by t r e a t - 1 ng subpopul a t i ons separate ly t h e under ly ing soci a1 processes may be more r e a l i s t i c a l l y portrayed.

The r e s u l t s presented i n Table 11 use d i f f e r e n t combinations o f va r i ab les f o r each age group t o est imate t h e l o g i t s f o r r e s i d e n t i a l m o b i l i t y (11). These models do not f i t t h e observed data as w e l l as t h e prev ious model. On t h e o the r hand, t h e o v e r a l l f i t o f these models ( t h e p r o b a b i l i t y t h a t dev ia t i ons as l a r g e as those observed would occur i f t h e model were t h e t r u e model) i s w i t h i n t h e range normal ly regarded as a good f i t . Moreover t a k i n g away any of t h e var iab les i n t h e models leads t o a s i g n i f i c a n t l y worse f i t , w h i l e t h e f i t may be improved on ly by adding i n t e r a c t i o n terms t h a t do no t l end themselves t o meaningful i n t e r p r e t a t i o n .

! As presented i n Table 11, however, t h e models suggest a number o f i n t e r p r e t - ab le r e s u l t s concerning t h e pa t te rns o f r e s i d e n t i a l m o b i l i t y among one-person households i n each age group. To begin wi th, t h e decreasing absolute value of t h e constant term as one goes from t h e o ldes t t o t h e youngest age group represents t h e successively i ncreasi ny p r o p o r t i on o f mobi le persons. The t r a n s i t i o n from a one-person households t o a mu1 t i -person household, though common i n t h e popu la t i on a t hand, i s re1 a t i ve ly r a r e i n t h e o ldes t age group and among members o f our sample was never associated w i t h r e s i d e n t i a l m o b i l i t y . Thus t h e r e i s no th ing t o be gained by i n c l u d i n g t h i s t ype o f change i n house- h o l d composit ion i n t h e model f o r t h e o ldes t age group. I n t h e two younger age groups, on t h e o ther hand, t h e t r a n s i t i o n t o a mul t i -person household i s c l e a r l y p o s i t i v e l y re1 ated t o r e s i d e n t i a1 mobi 1 i t y . To consider t h e possi b- i l i t y t h a t m o b i l i t y among t h e o lde r group o f persons cou ld represent t h e lagged response t o a change i n household composit ion p r i o r t o t h e s t a r t of t h e observat i on p e r i od, a dummy v a r i a b l e was constructed i n d i c a t i ng persons known t o have been widowed o r d ivorced i n t h e prev ious t h r e e years. However, t h e r e was no c l e a r c u t r e l a t i o n s h i p between t h i s v a r i a b l e and r e s i d e n t i a l mobi 1 i t y and i t was not inc luded i n t h e model.

The c o e f f i c i e n t s f o r t h e number o f rooms i n t h e dwe l l i ng a t T1 a re a good example o f t h e con t ras t between t h e age groups. I n t h e two youngest age groups a small dwel li ng i s posi ti ve l y associated w i t h r e s i den t i a1 mobi 1 i ty . As discussed above, i n t h e younger age groups r e s i d e n t i a1 m o b i l i t y i s i n t h e d i r e c t i o n o f l a r g e r dwel l ings and i n t h e o ldes t age group toward smal le r dwel l ings. Accordingly, i n t h e o ldes t age group we f i n d a negat ive c o e f f i c i - en t f o r persons i n t h e smal lest dwe l l ings ( these persons have a l ready a t t a i n e d t h e des i red dwe l l i ng s i z e ) and a p o s i t i v e c o e f f i c i e n t f o r persons i n t h e next l a r g e s t s i z e dwel l ing. Th is i s t h e group o f persons moving i n t o smal le r quarter. The c o e f f i c i e n t f o r three-room dwel l ings i s negat ive, as i s t h e coef f i c i ent f o r t h e v a r i ab le i n d i c a t i ng o l d e r one-person households 1 i v i ng i n one o r two f a m i l y houses as opposed t o m u l t i - u n i t bu i ld ings . Presumably persons i n t h e l a r g e s t o f dwe l l ings a re r e l a t i v e l y immobile because they have t h e f i n a n c i a l resources t o cont inue t o occupy such a dwel l ing. The i n c l u s i o n o f monthly household income i n t h e model f a i l e d t o capture t h i s r e l a t i o n s h i p , bu t t h i s i s a poor measure o f f i n a n c i a l resources f o r persons

--------- (11) The number o f rooms occupied a t t h e s t a r t o f t h e observat ion p e r i o d i s t h e

o n l y v a r i a b l e comnon t o a l l models and i n each case t h i s v a r i a b l e i s coded somewhat d i f f e r e n t l y : f o r t h e o ldes t age group we d i s t i n g u i s h between 1 room, 2 room, 3 room and 4 room, o r l a r g e r (omi t ted category) dwel l ings; and f o r t h e youngest age group between 2 room o r smal le r and 3 room o r 1 arger (omi t ted category) dwel l ings. For t h e midd le age group between 1 room and 2 room o r l a r g e r (omi t ted category) dwel l ings.

i n t h i s age group. A v a r i a b l e f o r whether t h e d w e l l i n g was owned o r ren ted a l s o f a i l e d t o improve t h e model, p r i m a r i l y because i n t h i s age group nea r l y a l l p roper ty owners are owners of l a r g e dwel l ings and thus t h i s v a r i a b l e i s redundant. Subst i t u t i ng proper ty ownership f o r e i t h e r o r both i n d i c a t o r s of dwe l l i ng s i z e on ly decreases t h e o v e r a l l f i t o f t h e model.

For persons i n t h e youngest age group we added a v a r i a b l e t o consider our hypot hesi s t h a t mobi 1 i t y among t h e youngest one-person households i s concen- t r a t e d among those persons who have not y e t s e t t l e d i n t o occupat ional l i f e .

I As expected we found a p o s i t i v e re1 a t i onsh i p between r e s i d e n t i a1 mobi 1 i t y and a dummy v a r i a b l e i n d i c a t i n g t h a t a person was s t i l l i n school o r occupat ional t r a i n i n g a t t h e s t a r t o f t h e observat ion period. Moreover t h e o v e r a l l fit f o r our model f o r persons i n t h e youngest age group on ly became t o l e r a b l e when we

I added an i n t e r a c t i on term represent i ng persons s t i 11 i n school o r occupat ional t r a i n i n g a t T 1 and l i v i n g i n 2 room o r smal le r dwel l ings. Here one can conclude t h a t t h e r e l a t i o n s h i p w i t h r e s i d e n t i a l m o b i l i t y decreases when a person has

I a1 ready acqui red a more desi r a b l e d w e l l i n g ( i n t h i s age group more desi r a b l e means l a r g e r and not smal ler as i s t h e case w i t h o l d e r persons) be fore complet- i n s h i s o r her education.

I

Constant -1.516 ( 132)

Change from one-person household to multi- .. person household 9

Number of rooms in dwelling

1

1 or 2 family house

In cchool or vocational training at T1

In cchool or vocational training in dwelling with I 2 rooms at T1

Likelihood ratio chi-square df Probability

. . Based on waves 1-3 of the SOP. n - 997 (etandard errors) I Table 11

Logit Llodalm with Dapandent Variable: novad/Rmained a t 8-0 Addrarm 8afirata nodal for Eaah age Group

IItIItItItIIIIIIIII

29

4 CONCLUSIONS

The f indings presented above demonstrate how the growth in the number ofone-person households can be interpreted by examining the social processesthat are the const i tuent elements of th is growth. In th is context thecomparison betneen the FRG and the US is part icular ly instruct ive. Inthe t ime per iod considered relat ively nore new one-person households wereformed in the US than in the FRG. However the increase in the number ofone-person households was relat ively greater in the FRG than in the US,because relat ively more one-person households were being dissolved in theUS. Previous res-earch using'panel data has emphasized an important pointin conjunct ion wi th the study of unemployment and poverty: the proport ionof persons in a g ' iven state depends both on the rate at which people_enter.and ' leave the s t i te . Th is p r inc ip le a lso app l ies to type o f househo lds ands tages o f fami ly l i fe .

Typical ly one thinks of a young one-person household as a temporary l iv !n9airangement between the parental household and the formation of a householdwith a partner or spouse. However, our f inding show that a great number ofone-perSon househol i ls are formed, part iculary in the Uni ted States'_throughthe

-d ' issolut i on of a marr i age or partnership. Soci o logists of the fami ly

have descr ibed changing patterns of l iv ing-together, marr iage' d ivorce andremarr iage as "ser i i l mohogamy". The research presented above indicatesthat one-person households-of len serve as a stepping stone in the t ransi t ionfrom one partnership to the next. 0n the other hand, the t ransi t ion f rom. aone-persoh househotb to a mult i -person household of ten does not lead to theformat ion of a partnership but tb a reintegrat ion in the parental household.Our resul ts indicate that ' th is t rend, noted by others in the US (Heer, Hodgeand Felson 1985), is of far less importance in the FRG. The maior i ty oft ransi t ions to mult i -person households in the FRG lead to the format ionof a household wi th a partner--more of ten wi thout than with a marr iagecerti f i cate.

The descr ipt ion of resident ia l mobi l i ty among one-person householdsi l lustratei an important aspect of the process of resident ia l mobi l i . ty_- -in general . In the course bf tne' i r l ives persons in the US change dwel l ingson [he average far more frequently than those in the FRG, wherq tlte averageperson thi rti years of age dr youhger has moved three times and those betweentnirty and sixiy-five only foui t imes. Greater residentlal mobil ity andmore irequent changes in household composition and family status are notindependdnt of one-another, for these events ln fact often accompany oneanother. The low rate of residential mobil ity among young one-personhouseholds wi thout changes in household composl t ion in the US providesstrong support for th is argument. In the absence of_changes in householdcompoiit ibh trousing conceris and, for the youngest of persons, 999!Pationalplacement are the most lmportant determinants of residential nobil ity.

As the final section of the paper emphasizes, different t ltqes of one-personhouseholds exhibit different- pitterni of residential_mobil ity. . l lhi le rteiound no relat ionship between' the t ransl t ion to a mult i -person household andresident ia l nrobi l i ty in the oldest age group, in the two younger age grouPsa change ln househoid composl t ion wai posi t ively assoclated wlth a change Inhouseh6ld conposition exhibited by difierent types of households is necessaryto irprove thb accuracy of models-of residential mobil lty. To the extent

t h a t housing and l a b o r market p o l i c i e s a re aimed a t t h e a l l a c a t i o n of persons t o s p e c i f i c areas t h e housing needs of young one-person households, t h e most mobi le type o f household, a re worthy o f spec ia l a t t en t i on . Ghviously t h i s imp l i es a s u f f i c i e n t number of adequate and a f fo rdab le dwel l ings i n t h e appropr ia te loca t ions . Less obviously, however, a t t e n t i o ~ m s t a l so be p a i d t o t h e fu tu re housing needs o f such persons; f o r many o f $#ese persons l i v i n g alone i s only a temporary s i t u a t i o n , fo l lowed by marr iage o r t h e formation of a household w i t h an unre la ted person. Nothing i s gained if persons a re success fu l l y a t t r a c t e d t o an area, bu t they arp t h e n fo rced t o move elsewhere because they a re unable t o f i n d appropr ia te housing f o r a mu 1 t i -person household.

The f ind ings presented above, i n pa r t , r e s t on a r e l a t i v e l y smal l number of observations. However they prov ide t h e fmmdat ion f o r mere exact models of changes o f household composi t i on i n v o l v i n g me-perscm households and th-e re1 ated processes o f r e s i den t i a1 mob i l f t y . For tunate ly t h e SIPP and SOP pane1 p r o j e c t s are such t h a t these p re l im ina ry r e s u l t s may bnr b u i l t upon and research i n t h i s d i r e c t i o n may be continued. The t o p i c a l modules i n t h e 1984 S I P P panel concerning m ig ra t i on and mar$$al h isaary promise a weal th o f a d d i t i o n a l data. The ongoing nature o f t h e SOT p ~ m e l --data from t h e fou r th wave w i l l soon be a v a i l a b l e and t h e f % f t h wave i s c u r r e n t l y b e i n g co1 1 ected--wi 1 1 a1 1 e v i a te some o f t h e sample s i z e p r o b l e m s s o c i ated defith t h e study of changes i n household composit ion fn t h e FRG, W e r t h e course of t h e panel, cases o f t h e fo rmat ion of one-permin househ@I&s and t h e trams- i t i on from one-person households t o mu1 t i -person househol& accumulate. Assuming t h a t t h e processes i nvo l ved have themselves no t changed, obsewa- t i ons from d i f f e r e n t p o i n t s ifl t ime be romb i ned -al4+wing f o r more d e t a i l e d analys is . F i n a l l y , we hope our work w i l l encourage o thers t o undertake comparative panel research. I n d i v idua l -o r ien ted l o n g i t u d i n a l data such as t h a t o f f e r e d by S I P P and SOP, and t h e sof tware and hardware necessary t o work w i t h such data, have on ly recen t l y become avai lab le. By drawing a t t e n t i o n t o t h e s i m i l a r i t i e s and d i f f e r e n c e s between processes i n more than one soc ie ty , a comparative approach can be a great a i d i n makang sense o f t h e weal th o f i n fo rma t ion o f f e r e d by data o f t h i s type.

REFERENCES

DUNCAN, Greg J. (1985), " A Framework f o r Track ing Fami ly Re la t ionsh ips Over Time", Journa l o f Economi c and Soci a1 Measurement, 13, 237-244.

HANEFELD, Ute (1984), "The German Socio-Economi c Panel ", 1984 Proceedings of t h e Soc ia l S t a t i s t i c s Section, Washington, DC, 117-123.

HEER, David M., HODGE, Robert W. and FELSON Marcus (1985), "The C l u t t e r e d Nest: Evidence t h a t Young Adu l ts a re More L i k e l y t o L i v e a t Home Now than i n t h e Recent Past", Sociology and Soci a1 Research, 436-441.

HUGHES, James W. (1980), "Demographic and Economic Change: Imp1 i c a t i o n s f o r Mobi l i t y " , 59-78 i n CLARK, W.A.V,, and MOORE, ERIC G. (eds.), Res iden t i a l Mobi 1 i t y and Publ i c Pol i cy , Bever ly Hi 11 s, London: SAGE Publ i c a t i ons.

KALTON, Graham, and LEPKOWSKI, James ( l 985 ) , "Fo l low ing Rules i n SIPP", Journal o f Economi c and Soci a1 Measurement, 13, 319-330.

KALTON, Graham (1985), "Hand1 i ng Wave Nonresponse i n Long i tud ina l Surveys, unpubl i shed manuscript, Uni v e r s i t y o f M i c h i gan.

KNOKE, David, and BURKE, Peter J. (1980), Log-Linear Models, Bever ly Hi 11s: SAGE Publ i c a t i ons.

LAHMANN, Herber t (1988), "Wohnen", for thcomi ng i n H.J. KRUPP and JUERGEN SCHUPP (eds. ) , Lebensl agen i m Wandel : Daten 1987, Frankfurt INew York: Campus Ver l ag.

LITTLE, R., and DAVID, M. (1983), "Weight ing Adjustments f o r Non-response i n Panel Surveys", Bureau o f t h e Census Working Paper.

McMILLEN, David B. and HERRIOT, Roger (1985), "Toward a Long i tud ian l De f i n i t i on o f Households", Journal o f Economic and Soci a1 Measurement, 13, 349-360.

QUIGLEY, John M. (1987), " I n t e r e s t Rate Var ia t ions , Mortgage Prepayments and Household Mobi 1 i t y " , The Review of Economics and S t a t i s t i c s , 53, 636-643.

RENDTEL, Ul r i c h (1987), "Methodische Konzepte f u e r d i e Hochrechnung von Panel- Daten", V i e r t e l j a h r s h e f t e z u r W i r t schaf ts fo rschuny , Heft 4.

R I M A , Annemarie, and van WISSEN, L.J.G. (1987), A Dynamic Model o f Household Re loca t i on, Amsterdam: Free U n i v e r s i t y Press.

ROUSSEL, Loui s ( 1 983), "Les menages d l une personne: 1 ' evo l u t i on recente" , Populat ion, 38, 995-1016.

SPIEGEL, E r i k a (1986), Neue Haushaltstypen, Frankfurt /New York: Campus Verlag.

STATISTISCHES BUNDESAMT ( 1 987), Fachseri e 1: Bevoel kerung und Erwerbs tae t i gkei t , Reihe 3: Haushalte und Fami l i en 1985, S t u t t g a r t , Mainz: Kohlhammer.

U.S. BUREAU OF THE CENSUS (1984), Current Popu la t ion Reports, Ser ies P-70, No. 1, "Economic C h a r a c t e r i s t i c s o f Households i n t h e Un i ted States: T h i r d Quar te r 1983", Washington, DC: U.S. Government P r i n t i n g Office.

U. S. BUREAU OF THE CENSUS (1985), Cur ren t Popul a t i on Reports, Se r i es P-70, No. 5, "Economic C h p r a c t e r i s t i c s of Households i n t h e Un i ted States: T h i r d Quar te r 1984", Washington, DC: U.S. Government P r i n t i n g Office.

U.S. BUREAU OF THE CENSUS (1987a), S t a t i s t i c a l Abs t rac t o f t h e Un i ted Sta tes 1987, Washington, DC: U.S Government P r i n t i n g O f f i ce .

U.S. BUREAU OF THE CENSUS (1987b), Survey o f Income and Program P a r t i c i p a t i o n User 's Guide, Washington, DC: U.S. Government P r i n t i n g Of f i ce .

WAGNER, Michael (1987), "Raeuml i c h e Mobi 1 i t a e t i m Lebensverlauf", for thcoming PhD-Di s s e r t a t i on a t t h e Free Uni ve rs i t y o f B e r l in .

WITTE, James (1988), "Haushalt und Fami l ie " , for thcoming i n H.J. KRUPP and JUERGEN SCHUPP (eds. ) , Lebenslagen i m Wandel : Daten 1987, F rank fu r t / New York: Campus Verlag.